MLflow Tracking Server Key
Service Name: MLflow
Service Description: MLflow is an open-source platform for managing the end-to- end machine learning lifecycle. It provides tools for tracking experiments, packaging code into reproducible runs, and sharing and deploying models.
Service Address: https://mlflow.org/
Validation Type: API Auth
IP Allow list: IP restrictions can be configured at the server level depending on deployment configuration.
Secret Access Scope: Grants access to MLflow tracking server for logging parameters, metrics, artifacts, and managing experiments.
Secret Revokement URL: Does not exist (managed through server configuration)
Secret Example: mlflow-tracking-token-a1b2c3d4e5f6g7h8i9j0
Suspicious Activity Investigation Instructions:
- Review MLflow tracking server logs for unusual access patterns or experiment modifications
- Check for unexpected experiment creations or model registrations
- Monitor for unusual artifact uploads or downloads
- Examine tracking server access logs for unauthorized IP addresses
- Look for abnormal usage patterns such as high-volume requests or off-hours activity
Mitigation Instructions:
-
Rotate the tracking server authentication token by updating the server configuration
-
Update authentication settings in the MLflow tracking server configuration
- Remove the compromised token from authorized credentials
- Implement more restrictive access controls on the MLflow tracking server
- Consider implementing additional authentication mechanisms such as OAuth or
LDAP
- Review and update network access controls to limit server accessibility
- Audit all experiments and models for unauthorized changes